The anterior cruciate ligament (ACL) of the knee joint can be divided anatomically into an anteromedial and posterolateral bundle. Yet, most current ACL reconstruction techniques focus on replacing only the anteromedial bundle. Biomechanical studies have revealed that reconstruction using a single-bundle graft is successful in limiting the anterior tibial translation, but insufficient in controlling excessive tibial rotation. As persistent abnormal knee kinematics are a putative factor in the degeneration of cartilage following ACL injury, interest has arisen in a double-bundle technique in an attempt to replicate more closely the natural anatomy and function of both ACL bundles. Recent in-vitro studies have shown that the anatomic reconstruction of the ACL using two graft bundles indeed produces a better biomechanical outcome, especially close to knee extension and during rotatory loads. However, it remains unclear whether these promising in-vitro results obtained from cadaver studies translate into improved long-term clinical outcome. Although an increasing number of surgeons embrace the double-bundle ACL reconstruction, the question needs to be addressed whether the more complicated surgical technique improves the in-vivo kinematics of the knee joint. To answer this question, fifty-six patients with a unilateral ACL injury will be prospectively randomized into two groups, a single-bundle hamstring group, and a double-bundle hamstring group. All the reconstructions will be performed by the same surgeon. Both knees of each patient will be imaged with a magnetic resonance scanner to create three-dimensional (3D) knee models. Next, each knee will be imaged with two fluoroscopes prior to the ACL reconstruction, and six months and twelve months after the ACL reconstruction during a series of dynamic activities. The fluoroscopic images and corresponding 3D knee model will be introduced into solid modeling software. The 3D models will be manipulated in six degrees-of-freedom until the projections of the models match the outlines of the fluoroscopic images. The 3D knee models will thereby reproduce the in-vivo knee positions, and the tibiofemoral kinematics of the single-bundle and double-bundle reconstructed knees will be compared. ACL injury most affects patients under thirty years of age, and is clinically associated with an increased incidence, earlier onset, and faster progression of knee osteoarthritis. Even a small increment of improvement in function would be valuable for this active population. However, we still need to determine whether drilling additional tunnels in the knee joint is an advance in-vivo, or a step in the wrong direction.